Deep learning–based statistical robustness evaluation of intensity-modulated proton therapy for head and neck cancer

<i>Objective</i>. Previous methods for robustness evaluation rely on dose calculation for a number of uncertainty scenarios, which either fails to provide statistical meaning when the number is too small (e.g., ∼8) or becomes unfeasible in daily clinical practice when the number is sufficiently large (e.g., >100). Our proposed deep learning (DL)-based method addressed this issue by avoiding the intermediate dose calculation step and instead directly predicting the percentile dose distribution from the nominal dose distribution using a DL model. In this study, we sought to validate this DL-based statistical robustness evaluation method for efficient and accurate robustness quantification in head and neck (H&N) intensity-modulated proton therapy with diverse beam configurations and multifield optimization.<i>Approach</i>. A dense, dilated 3D U-net was trained to predict the 5th and 95th percentile dose distributions of uncertainty scenarios using the nominal dose and planning CT images. The data set comprised proton therapy plans for 582 H&N cancer patients. Ground truth percentile values were estimated for each patient through 600 dose recalculations, representing randomly sampled uncertainty scenarios. The comprehensive comparisons of different models were conducted for H&N cancer patients, considering those with and without a beam mask and diverse beam configurations, including varying beam angles, couch angles, and beam numbers. The performance of our model trained based on a mixture of patients with H&N and prostate cancer was also assessed in contrast with models trained based on data specific for patients with cancer at either site.<i>Results</i>. The DL-based model's predictions of percentile dose distributions exhibited excellent agreement with the ground truth dose distributions. The average gamma index with 2 mm<i>/</i>2%, consistently exceeded 97% for both 5th and 95th percentile dose volumes. Mean dose-volume histogram error analysis revealed that predictions from the combined training set yielded mean errors and standard deviations that were generally similar to those in the specific patient training data sets.<i>Significance</i>. Our proposed DL-based method for evaluation of the robustness of proton therapy plans provides precise, rapid predictions of percentile dose for a given confidence level regardless of the beam arrangement and cancer site. This versatility positions our model as a valuable tool for evaluating the robustness of proton therapy across various cancer sites.